USING DRONES TO UNDERSTAND FLUVIAL SYSTEMS: LESSONS LEARNED FROM A YEAR OF FLYING ALONG THE RIO GRANDE

Small Unmanned Aircraft Systems (sUAS or drones) outfitted with digital cameras have recently emerged as crucial data collection tools for understanding fluvial systems and managing watersheds and rivers. Images captured at low-altitude (less than 400 ft above ground level) are processed with Structure-from-Motion (SfM) photogrammetry software to generate orthomosaic maps and digital surface models. The data and map products are of similar accuracy, but often at much higher-resolution (typically less than 5 cm) than other established remote sensing and photogrammetry techniques. In the spring of 2017, the New Mexico Bureau of Geology initiated a sUAS program to begin assessing the role that drones can play in monitoring surficial processes. Extensive testing using repeat flights of two field areas, an ephemeral stream and a recently burned section of the Rio Grande floodplain, is establishing best practices and identifying some limitations of the sUAS-SfM technique. Important considerations that will be discussed in this presentation include the relationship between desired accuracy and the role of ground control points, the type of sUAS platform (rotary vs. fixed wing) needed given the size of the area to be mapped, and developing multifaceted flight patterns that both maintain visual line of sight and generate the desired resolution and accuracy. In most scenarios, the sUAS-SfM technique produces cm to dm-scale relative accuracy elevation models, but referencing SfM maps to ground control points measured with RTK-GPS systems ensures cm-scale absolute accuracy. Our initial findings suggest that most topographic features can be accurately imaged with the sUAS-SfM technique. However, capturing elevation data in densely forested areas and vegetated micro-topography may be challenging for the technique. Certain environmental conditions, such as wind and lighting conditions, can impact the map quality. Performing flights close to solar noon or on overcasts days limits the impacts of shadows on the elevation data. In addition to high-resolution orthomosaic and digital surface models, examples will be presented that show how RGB photos can be processed to understand the greenness of vegetation and bedforms, which for some studies may be useful a component of characterizing the fluvial system.